Stroke study shows variation in care

Benchmark to improve outcomes and lower costs

Don't be surprised if benchmarking data show your hospital is handling stroke patients in a markedly different way than your peers. A recent study of nearly 1,000 patients at 36 academic institutions proved that no two of those hospitals treat acute ischemic stroke patients exactly alike. The hospitals' average lengths of stay for stroke patients, for example, varied from fewer than three days to more than 21 days, the study showed.

That wide variation could hinder hospitals' efforts to improve stroke patients' care and to reduce costs. If all hospitals were to follow a best practices approach, the overall impact could be huge: 740,000 people have strokes each year, and the total annual cost of treating them exceeds $40 billion, says Lawrence M. Brass, MD, professor of neurology at Yale University in New Haven, CT. Brass was one of the researchers involved in the study, which was sponsored by the University Health System Consortium (UHC) in Oak Brook, IL. "Stroke is the third leading cause of death and the leading cause of serious disability in adults, but there has not been a huge amount of benchmarking of stroke care despite its importance," he says.

The UHC conducts benchmarking projects to help its members identify their own care processes and how these vary with other institutions, says Jean Livingston, RN, MS, director of clinical process improvement at the UHC. The consortium is an alliance of academic health centers with more than 80 national and international members, which have been conducting benchmarking and other studies since 1994. "Hopefully, people are using the information and feedback that results from analyzing this data to identify their own opportunities for improvement," she says.

The UHC's stroke study has shed some light on treatment, Brass says, with these three surprising results:

1. There's a wide variation in how stroke patients are treated. "One would think that at the best medical centers, there would be some consensus in how to do things, but we found instead a wide variation," Brass says.

2. Hospitals are underutilizing the methods and medicines that work. For example, aspirin and other anticoagulants are a cheap and simple way to prevent myocardial infarction and also might prevent strokes, but the number of patients given aspirin at discharge from the hospital ranged from 27% to 80%, with a median of 45%, says Judith H. Lichtman, PhD, MPH, associate research scientist in the department of neurology at the Yale school of medicine. Lichtman also is an epidemiologist at the Center for Outcomes Research and Evaluation at Yale-New Haven Hospital. (See summary of stroke benchmarking study, p. 59.)

Livingston says the study shows hospitals how many patients are receiving aspirin or other oral anticoagulants so they can use the information to see whether the institution is following its own or national guidelines. "Very often, it is a situation where people think they are following a guideline, but they are not because of a systems issue or communications problem," she explains.

3. Hospitals are overutilizing therapies that research has shown do not help acute stroke patients. High blood pressure, for example, puts a person at risk for a stroke, and medication to lower blood pressure should be used to prevent strokes. But there is an important exception: National guidelines advise against using blood pressure medication in patients who have just had an acute stroke, Brass says. After an acute stroke, the patient's body reacts by increasing blood pressure to try to get more blood in the brain. "The standard guidelines all suggest that very few patients need treatment in the acute stage, except in the event of a heart attack."

The study showed that 57% of the stroke patients were given antihypertensive medications after admission, despite the contraindications. Of the patients receiving hypertensive medications, 45.5% were given new medications, and 54.5% were continued on medications they had been prescribed before admission.

"These are well-established guidelines," Brass says. "It's not the time to continue patients on blood pressure medication." (See peer review guidelines for treating stroke patients, p. 59.)

The consortium will sponsor a follow-up study to see if the 36 academic institutions have had any improvement in outcomes in the two years since the first study was completed, Livingston says.

"I believe all of our participants are using the information now to improve their process of care and to improve their outcomes," Livingston says.

The next phase will examine whether the hospitals have achieved shorter lengths of stay. It also will look at where patients have been discharged. "We need to take into consideration whether they were discharged into home care or into a long-term facility or that kind of thing," she adds.

These types of benchmarking studies are not meant to be punitive or critical of the facilities that have poorer outcomes, Lichtman explains. Instead, they may help the health care industry establish clinical guidelines that could reduce lengths of stay and costs, she says, adding that the cost information from the first study has not been released.

"One of the strengths of this data set is you're looking at how things are done across a number of hospitals," Lichtman says. Another strength is that the study looks at 30 consecutive patients. "You're not saying, `Let's take the five patients who look the best.' You're getting the next 30 who come in the door."